Re: [systemd-devel] systemd mount point disappearing on underlying device loss?

2017-04-03 Thread Andrei Borzenkov 
03.04.2017 10:52, Michal Sekletar пишет:
> On Sat, Apr 1, 2017 at 6:46 AM, Andrei Borzenkov  wrote:
>> As far as I can tell, current systemd should behave as you want since
>> commit 9d06297. If you want this commit to be backported, you need to
>> contact your distribution.
> 
> 9d06297 changed behavior only for mount units created from
> /proc/mountinfo notifications. For normal mount units there is still
> BindsTo dependency to respective device unit.
> 

Thanks, missed that.

Then, it does not appear any way to bypass it.
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/systemd-devel

Re: [systemd-devel] Renaming interfaces with systemd.link

2017-04-03 Thread Chris Trobridge 
> Note that "networkctl status ..." will actually show you the link file
> applied to a specific interface. What does it show for yours?
> 
> Lennart

It shows "/usr/lib/systemd/network/99-default.link", despite the presence of my 
earlier named 10-eth0.link file.

The initramfs for Fedora contains "/usr/lib/systemd/network/99-default.link".

I just tried this on my Yocto box, which has no initramfs and the renaming is 
applied to eth0 successfully.

I have also renamed other interfaces with Fedora, eg a virtual bridge, with a 
.link file.

So my thought is that it is being renamed while initramfs is being used and 
then cannot be renamed again (as stated on the man page).  This would tie with 
the Debian issue I linked in my initial email 
(https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=793374).

Cheers,
Chris
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/systemd-devel

Re: [systemd-devel] Renaming interfaces with systemd.link

2017-04-03 Thread Lennart Poettering 
On Sat, 01.04.17 20:45, Chris Trobridge (christrobri...@hotmail.com) wrote:

> $cat /etc/systemd/network/10-eth0.link
> [Match]
> #OriginalName=eth0
> MACAddress=08:00:27:57:f6:40
> 
> [Link]
> Name=eno1
> 
> $ dmesg | grep eth0
> [1.636915] e1000 :00:03.0 eth0: (PCI:33MHz:32-bit) 08:00:27:57:f6:40
> [1.636920] e1000 :00:03.0 eth0: Intel(R) PRO/1000 Network Connection
> [1.637818] e1000 :00:03.0 enp0s3: renamed from eth0
> 
> For these tests I have been using Fedora 25 in VirtualBox.

Note that "networkctl status ..." will actually show you the link file
applied to a specific interface. What does it show for yours?

Lennart

-- 
Lennart Poettering, Red Hat
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/systemd-devel

Re: [systemd-devel] systemd mount point disappearing on underlying device loss?

2017-04-03 Thread Michal Sekletar 
On Sat, Apr 1, 2017 at 6:46 AM, Andrei Borzenkov  wrote:
> As far as I can tell, current systemd should behave as you want since
> commit 9d06297. If you want this commit to be backported, you need to
> contact your distribution.

9d06297 changed behavior only for mount units created from
/proc/mountinfo notifications. For normal mount units there is still
BindsTo dependency to respective device unit.

Also commit 9d06297 was backported to CentOS 7.2.

Michal
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/systemd-devel